AMD K9 coming 2Q '05

guesswhosback · Oct 16, 2003

http://msnbc-cnet.com.com/2100-1006_3-5091845.html?part=msnbc-cnet&tag=alert&form=feed&subj=cnetnews

Snorkius · Oct 16, 2003

<A HREF="http://msnbc-cnet.com.com/2100-1006_3-5091845.html?part=msnbc-cnet&tag=alert&form=feed&subj=cnetnews" target="_new">http://msnbc-cnet.com.com/2100-1006_3-5091845.html?part=msnbc-cnet&tag=alert&form=feed&subj=cnetnews</A>


I will not add another word.
Horace

Parhelia · Oct 16, 2003

impressive indeed, hehe, the double core

eden · Oct 16, 2003

AMD is also looking at adding threading to future chips. Simultaneous multithreading essentially enables a chip to run two applications, or two "threads" of the same application, at the same time--thereby reducing the time it requires to complete a task. Threading, however, does not provide as much of an improvement to overall performance as multicore technology does, according to Weber.

Looks like AMD knows Intel was right with HT. This better come in it.

K9 has to be a total recreation. I hope so, it'd likely succeed.

--
<A HREF="http://www.lochel.com/THGC/album.html" target="_new">This just in, over 56 no-lifers have their pics up on THGC's Photo Album! </A> :lol:

slvr_phoenix · Oct 16, 2003

Looks like AMD knows Intel was right with HT. This better come in it.

K9 has to be a total recreation. I hope so, it'd likely succeed.

Man, I'd sooooooooooo love to see AMD come out with a complete redesign for K9 that includes HT. That'd rock. Hopefully it'd also be designed more around out of order execution as well because if AMD could in any way better utilize their superior FPU that'd be really awesome.

K8, while cool in some respects, was really a big disappointment too. It's just too close to K7's architecture to really get excited about. It'd be like if Intel had stuck to the P3 all these years and had only now just released the PentiumM as the first P4. :\ Depressing. Even if the performance is improved and there are some neat tricks, the architecture change just isn't awe inspiring.

So hopefully K9 will come out quickly and be a totally new architecture.

And hey, wouldn't it be weird if Transmeta wrote AMD64 support into their Crusoe? :O Something to think about...

<pre><A HREF="http://ars.userfriendly.org/cartoons/?id=20030905" target="_new">People don't understand how hard being a dark god can be. - Hastur</A></pre>

HardWareBoss · Oct 16, 2003

But Hammer was suppose to be here in 2001. So when will K9 make it? I do not think it will be 2005.

slvr_phoenix · Oct 16, 2003

But Hammer was suppose to be here in 2001. So when will K9 make it? I do not think it will be 2005.

Possibly. But think of it this way: What if AMD had been trying to develop a completely new core all along?

Once the delays started to put pressure on them to release something, anything, they just put as much of their new technology as they had into a K7 core without a serious redesign and called it K8. And that would have bought them time to finish designing their originial development which they could then release as K9. So it's possible that the K9 core has actually been in development for several years already.

I'm not saying that's what happened. I mean who could possibly know? I'm just saying that I can see it as possible.

<pre><A HREF="http://ars.userfriendly.org/cartoons/?id=20030905" target="_new">People don't understand how hard being a dark god can be. - Hastur</A></pre>

HardWareBoss · Oct 16, 2003

Now that would be something if that's what the are doing. K9 looks interesting. But so does Tejas. Plus the next one after that.

pitsi · Oct 16, 2003

I think the Athlon 64 will have an important role on whether the K9 will be here on time or not. If it is succesful and AMD manages to make profit out of the K8, and finally get some money on their hands, then hopefully they can invest as much as it is needed on research.

Bosse · Oct 16, 2003

Cyrix will relese a 56 ghz cpu later this fall !!!

beware !!!

AMD 2500+ @ 3200+ - Volcano 9 - A7N8X/DLX - Corsair XMS TWINX 3200LL 1024MB - GF 3 TI-200 - 2x WD Raptor RAID 0 - Maxtor DiamondMax Plus9 120Gb 7200rpm Special Edition - Hiper 420w.

pied_piper2004 · Oct 17, 2003

Is AMD the only processor with an integrated memory controller? Does that get classed as a trick?

Maybe not a complete redesign, but truly an effective redesign at that, I'd be willing to bet this architecture will hold out pretty well... look at how great it is right out of the gates... and the scalability is truly inspiring.

If it isn’t broke don't fix it, as some say.

Snorkius · Oct 17, 2003

I bet Intel and AMD have already shat their pants.


I will not add another word.
Horace Edited by snorkius on 10/16/03 11:25 PM.

pied_piper2004 · Oct 17, 2003

It sounds like research may be done for the K9 - Judging from leaked info THE K9 may be multi-core and have HT, sounds like it doesn’t need anymore research.

I'm talking out of my ass here, but don't they "invent/design" processors today for something they may release 5 years from now.

eden · Oct 17, 2003

K8, while cool in some respects, was really a big disappointment too. It's just too close to K7's architecture to really get excited about. It'd be like if Intel had stuck to the P3 all these years and had only now just released the PentiumM as the first P4. :\ Depressing. Even if the performance is improved and there are some neat tricks, the architecture change just isn't awe inspiring.

So hopefully K9 will come out quickly and be a totally new architecture.

Heheh, we think alike Slvr.

It doesn't impress me that much. Being a freak of silicon beauty and features (silicon porn, now that's the way!), the K8 brought so little to dazzle me. The P7 core is still superior in its features, but fails at delivering what could have. I strongly suggest AMD goes for a new core, extended pipeline, and tons of features that drive away from the K7 core's components but keeps faithful to AMD's WIDE architecture, meaning lots of IPC. Except this time, it'll be LONG and WIDE. Which could be very neat to have a CPU with superior scalability, like P4, and with an IPC of the K8. It's almost too good eh.

--
<A HREF="http://www.lochel.com/THGC/album.html" target="_new">This just in, over 56 no-lifers have their pics up on THGC's Photo Album! </A> :lol:

eden · Oct 17, 2003

Maybe not a complete redesign, but truly an effective redesign at that, I'd be willing to bet this architecture will hold out pretty well... look at how great it is right out of the gates... and the scalability is truly inspiring

We're not commenting against the performance, since we know very well it's the end performance that matters, but we're looking at it technically. After 4 years of the core, they bring out another that's the same but with some added features. It's just not too exciting.
They never even covered the extra 2 stages for crying out loud. Why not? Who knows what they can do? Has anyone read anything on the capabilities of these "packing" stages?

Also, Intel took about 5 years to design the P7, and there were rumors a good while ago that K9 research had begun. Likely they did, just like the K7 was started ~1997. So 2005 isn't a very blurry date, it's quite possible. Now making sure the multicore works right and without problematic thermal control is gonna be a new challenge, and with SOI wafers costing already a lot, can we gamble at the potential cost PER chip containing 2 cores? And what the market price will be so that profit levels keep consistent with the current?

--
<A HREF="http://www.lochel.com/THGC/album.html" target="_new">This just in, over 56 no-lifers have their pics up on THGC's Photo Album! </A> :lol:

pied_piper2004 · Oct 17, 2003

Well regardless of the bells and whistles, it's a great achievement... the ends justify the means, in this case.

Would you be happier if we called the integrated mem controller “Speed bust technology”?

If need be (just like many car manufacturers) they build on something that is already tried and proven… (Look at Ferrari)…

Not to get excited, the mem controller and 64 bit extension with probably be more useful or prove to be more beneficial than Intel’s HT…

Why not get excited about those technologies.

And again, why dismiss the 64 bit, I guess will see.

-pied

eden · Oct 17, 2003

Because they're not actual core architecture components. If that were the case we could say SOI is a K7 core component.

No, the extra history counters on the K8 are. That's pretty neat. Not overly exciting, but anything added INSIDE the core, aside from 64-bit and SSE2 as they're both known things and not necessarily integral to the K7 core's functioning, is pretty good to me. Similarly while SSE3 could be cool, it also isn't much of a thing for Prescott to me, but that's by personal preference as I just like core detail.

Even though the K7 is tried and true, this isn't car analogy class and nor will it become. We compare them too often when in fact it's a different thing.
A car won't eventually run into such physical limits like CPUs when at extremely high clock speeds. The K8 will scale, but no one knows how far.

AMD can definitely make a K8 derivate but on a completely different way of functioning, but maintaing what made the K7 so powerful, the WIDE aspect. They can make a K8 derivate AND extend pipeline length to the P4's level if not more.
Would that hurt IPC? Yeah, a tad because of pipeline bubble and misprediction rates, but there are tons of ways to get around it. Besides, losing 10% IPC on an Athlon FX-64 performance level is nothing bad. It's still at least 50% better per clock than the P4.
But it can scale like mad. It could reach 3.2GHZ levels with the SOI technology behind it, and simply be superior in all aspects.
That's without counting the potential K9 components that AMD could invent, just like Intel's Trace Cache or Double-pumped ALUs, which could boost IPC as well.

All in all, the K9, if properly done, CAN truly be superior to the K7 and prove that it is indeed a step forward where the "If it ain't broke, don't fix it" saying has no place.

I don't want to see another K7 core, but that's purely out of personal opinion on the core technology.

--
<A HREF="http://www.lochel.com/THGC/album.html" target="_new">This just in, over 56 no-lifers have their pics up on THGC's Photo Album! </A> :lol:

flamethrower205 · Oct 17, 2003

Now I'm pissed @ AMD!! I wanted to upgrade, but now I won't and will wait till 2005 (Christmas) to get new comp. That means this 2000/2001 comp will be alive for 4/5 years!! HOLY [-peep-], I graduate in 2005! Never realized that! Sw000t! Vodka's on me

The one and only "Monstrous BULLgarian!"

eden · Oct 17, 2003

Too much emotions all at once, must PROCESS!

Damn you're hyper!
Study load reduction imminent!

--
<A HREF="http://www.lochel.com/THGC/album.html" target="_new">This just in, over 56 no-lifers have their pics up on THGC's Photo Album! </A> :lol:

flamethrower205 · Oct 17, 2003

lol
Yeah, I don't get enough physical relief now, and as much as I'd like to think jackin off takes away energy, it does't. I need to go lift some weights, but no time dammit!

The one and only "Monstrous BULLgarian!"

imgod2u · Oct 17, 2003

A thing to keep in mind is heat and power consumption, not to mention transistor count. There is a reason the P4 isn't a massive parallel execution beast like the K7, at high clockspeeds, it's power consumption is already really high. Can you imagine adding an extra issue port? An extra decoder? Twice the trace-cache issue rate? We're talking about some massive power consumptions.
I think the way of Banias definitely needs to be explored on the desktop side. More efficient utilization of resources available now, not just shoving more resources on a chip. HT is a good direction towards as well as the packing/unpacking stages on the K8.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.

Guest · Oct 17, 2003

> but we're looking at it technically. After 4 years of the
>core, they bring out another that's the same but with some
>added features.

Actually, Im quite in awe over those "some added features". The three most important ones, may well be among the best three ideas recently introduced into any MPU:
1) Hypertransport. glueless, cheap and fast SMP. the concept is really, really neat. I drooled all over those possible >4 way configurations showing different cpu's (yes, opterons and A64's) connected to each other, one cpu connected to the AGP tunnel, another to the PCI-X tunnel, each CPU having their own memory channel.. simple, but brilliant IMHO; makes designing SMP systems as easy as playing with lego

2) AMD64. I think we've discussed this already a bit

I'm a great fan of the concept.
3) integrated memory controller.

On the intel side, I'm a great admirer of hyperthreading. Also a beautiful technology, and pretty effective, at least for the P7. but other than that, there is little in the P7 core that I drool over. On the contrary, I think the most promising design from intel lately, is based on the more than 5 year old P6 core, ie banias, and soon Dothan. That really is the way forward IMHO, and I would love to see intel release a Dothan core that is slightly less optimized for low TDP, and gives us higher clockspeeds. At 2.4 GHz, Dothan could well be about the fastest cpu out there, bar none.

>and there were rumors a good while ago that K9 research had
>begun

Thats a given. A cpu design takes ~5 years easily; in fact I thought K8 and K9 where largely developped in parallel, and any idea's that could not be realized in time for K8, where handed to the K9 team.

>Now making sure the multicore works right and without
>problematic thermal control is gonna be a new challenge,

Easier than hyperthreading IMHO. K8 core is already multi core ready, it could be as "simple" as connecting both cores internally with hypertransport and adding cache coherency. The thermal density is another issue, but guestimates about opteron/A64 power consumption show its roughly half as hot as the upcomming prescott. So even without a die shrink, multicore should be feasable as a dual core will not use twice as much power as a single core, since a lot is shared (memory controller, HT, cache,..)

>and with SOI wafers costing already a lot, can we gamble at
>the potential cost PER chip containing 2 cores?

Actually, SOI wafers are cheaper than 300mm bulk wafers, and per CPU candidate, they are about even. Everyone always uses the 300mm wafers of intel as a cost argument, but right now, 200mm is cheaper per cpu. 300mm just helps getting more fab capacity until 300mm wafers come down in price.

>can we gamble at the potential cost PER chip containing 2
>cores?

If you take an opteron, roughly 60% of the die size is the 1 MB cache. The rest is for the memory controller, HT links and the core. So adding a second core while shrinking to a .09 process would roughly give you the same size of chip. IF AMD manages to gets in L2 caches a bit denser like intel, it should be a pretty small core really.

= The views stated herein are my personal views, and not necessarily the views of my wife. =

imgod2u · Oct 17, 2003

On the intel side, I'm a great admirer of hyperthreading. Also a beautiful technology, and pretty effective, at least for the P7. but other than that, there is little in the P7 core that I drool over. On the contrary, I think the most promising design from intel lately, is based on the more than 5 year old P6 core, ie banias, and soon Dothan. That really is the way forward IMHO, and I would love to see intel release a Dothan core that is slightly less optimized for low TDP, and gives us higher clockspeeds. At 2.4 GHz, Dothan could well be about the fastest cpu out there, bar none.

IMO, it is Banias/Dothan's low power consumption that allows it that performance edge. The current P7 core seems to be limited in performance mainly because of power requirements. It surprises me Intel hasn't considered this problem before it happened and put some of Banias's power saving features into Prescott.
And Banias/Dothan may have had some basis around the P6 core but it is definitely *not* a P6. In fact, I'd argue there are more significant changes from P6-Banias than from K7-K8.

The thermal density is another issue, but guestimates about opteron/A64 power consumption show its roughly half as hot as the upcomming prescott.

Erm, how do you figure? I know the typical power dissipation on the Athlon64 is about 50-60W but how do you know what Prescott's typical power dissipation will be? The TDP has been rumored at to be around 100W but realistic cases won't reach that. If you're going to use the TDP number, you may do well to look at the TDP on the Athlon64, the Athlon64 FX 51 at 2.0 GHz has a TDP of 89W. Hardly "half" of Prescott and definitely would be a problem in dual-core. Keep in mind that 90nm won't be a "wonder" solution to the heat problem. Gate leakage becomes a huge problem at 90nm and I doubt it would help power consumption all that much.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.

Guest · Oct 17, 2003

>If you're going to use the TDP number, you may do well to
>look at the TDP on the Athlon64, the Athlon64 FX 51 at 2.0
>GHz has a TDP of 89W.

No, this has been debated to death on so many sites.. the 89W figure is the maximum figure ANY Opteron will ever use. Its anyone's guess if that relates to a 2.4 GHz or a 3 GHz model. Aces' did an article on this recently, they stated AMD had not yet characterized those cpu's, and had given a broad, "safe" number for the oems, HS manufacturers, etc to live up, to be sure current motherboards, PSU's, and heatsinks would also allow future opteron based products. 89W does not apply to the FX51 or Opteron 246, it applies to entire hammer line, including future products.

Johan also did some tests with the Newisys server that allowed to measure power consumption, and his guess was the opterons he tested used no more than 50-60W each running burnK7. So half the TDP may be a bit of a stretch, but its safe to say the FX uses a lot less power than the upcoming prescotts.

<A HREF="http://www.aceshardware.com/read.jsp?id=55000268" target="_new">http://www.aceshardware.com/read.jsp?id=55000268</A>

= The views stated herein are my personal views, and not necessarily the views of my wife. =

CaptainNemo · Oct 17, 2003

I can't believe that no one has mentioned Doctor Who yet.

"Some mice have two buttons. Macintosh has one. So it's extremely difficult to push the wrong button." - Apple ad. circa 1984.

AMD K9 coming 2Q '05

Distinguished

Splendid

Distinguished

Champion

Splendid

Splendid

Splendid

Splendid

Distinguished

Distinguished

Distinguished

Splendid

Distinguished

Champion

Champion

Distinguished

Champion

Illustrious

Champion

Illustrious

Distinguished

Guest

Guest

Distinguished

Guest

Guest

Distinguished

Share this page